Aug 7, 1978 - A new method for studying membrane transport is presented. High resolution n.m.r. is used to measure the distribution of small molecules ...
Biochem. J. (1979) 180, 37-44 Printed in Great Britain
37
Application of Spin-Echo Nuclear Magnetic Resonance to Whole-Cell Systems MEMBRANE TRANSPORT
By KEVIN M. BRINDLE, FRANCIS F. BROWN, IAIN D. CAMPBELL, CHRISTOPH GRATHWOHL and PHILIP W. KUCHEL* Department of Biochemistry, University of Oxford, South Parks Road, Oxford OX1 3QU, U.K.
(Received 7 August 1978) A new method for studying membrane transport is presented. High resolution n.m.r. is used to measure the distribution of small molecules between the intracellular and extracellular compartments. The method uses spin-echo techniques and relies on a difference in the magnetic susceptibility of the media inside and outside of cells. It also provides simultaneous information on the metabolic status of the cell. The method is illustrated by a study of alanine and lactate transport in the human erythrocyte. N.m.r. is being used increasingly to study wholecell suspensions, especially by 31P n.m.r. (Moon & Richards, 1973; Navon et al., 1978). Such studies have shown that various metabolite concentrations and intracellular pH can be measured. Detailed metabolic information is also available from n.m.r. studies of the more sensitive and ubiquitous 'H nucleus (Brown et al., 1977), but such measurements rely on the spectral simplification achieved by applying spin-echo techniques (Campbell et al., 1975). ln an extension of our previous 'H spin-echo n.m.r. studies on whole cells, we demonstrate here a means of applying this technique so as to observe directly the membrane transport of any molecule that gives an observable n.m.r. signal. This information on transport can be obtained at the same time as the other information on the cell previously described. Studies of transport involve the determination of the distribution of the molecule of interest across the cell membrane as a function of time. Continuous monitoring procedures include: (a) light scattering to measure volume changes caused by the osmotic response to the changes in molecular distribution (Sen & Widdas, 1962; Sha'afi et al., 1967); (b) enzymic modification of specific extracellular species (Hertz & Barenholz, 1973). Separation procedures involve filtration or centrifugation and the determination of the distribution of a radioactively labelled molecule. Chemical methods are often necessary for stopping transport at exact time intervals (Eilam & Stein, 1974). N.m.r. studies of transport in cell suspensions depend on some method for distinguishing molecules inside and outside the cells. Transport of carboxylic acids into membrane vesicles has been studied by exploiting chemical-shift differences caused by pH * Present address: Division of Clinical Investigation, Faculty of Medicine, University of Newcastle, Newcastle, N.S.W., Australia.
Vol. 180
gradients across the membrane (Cramer & Prestegard, 1977). Differential relaxation on the two sides of the membrane has also been used to measure membrane permeability. For example, water transport into erythrocytes has been measured by 'IO n.m.r. (Shporer & Civan, 1975) and by 'H n.m.r. by adding Mn2+ ions to the' outside medium (Conlon & Outhred, 1972). Lanthanide ions have also been used in vesicle systems to cause rapid relaxation of the resonances of molecules outside the membrane (Hunt, 1975; Degani, 1978). Enzyme-catalysed exchange of 'H for 2H in a 1H n.m.r. experiment has been used to observe the efflux of L-alanine from erythrocytes (Brindle, 1978); the 2H labelling of alanine was achieved by addition of glutamatepyruvate transaminase (EC 2.6.1.2) to the extracellular space. These. methods are, however, difficult to apply to a wide va'riety of molecules and therefore : lack generality. In this paper we present an n.m.r. method which does appear to have the potential for widespread application. The transport of any molecule with an observable n.m.r. signal can be measured relatively easily. The method depends on the molecules on the two sides of the cell membrane giving rise to different intensities in the n.m.r. spectrum. This difference in signal intensities allows changes in the distribution of molecules to be measured as a function of time. The intensity differences arise because of magnetic susceptibility effects,and from the use of spin-echo techniques.
Experimental The erythrocytes from various subjects were prepared at room temperature from freshly drawn venous blood by washing once in 0.9% NaCl in 1H2O and either twice in 'H20 Krebs/Ringer buffer or four times in 2H20 Krebs/Ringer buffer (Krebs &
38 Henseleit, 1932) that was thoroughly gassed with 02/C02 (19:1). All solutions contained 10mMglucose. The samples were run on a Bruker 270 MHz FT n.m.r. spectrometer fitted with quatrature detection. Spectra were accumulated in the spinecho mode with a 90'-r-1800-r pulse sequence (900 pulse 15ps) where -r is the delay time, usually 60ms and with Is overall repetition rate as previously described (Brown et al., 1977). The accumulated decays (10-512 scans) were stored on disc by a computer-controlled automatic data acquisition routine, which also provided accurate timing for all rates of transport studied. On completion of an experiment, the haematocrit of the sample was measured on a Hawkesley microhaematocrit centrifuge. The addition of substrate species to the erythrocyte suspensions was made from iso-osmotic solutions. The samples were in 5 mm diameter tubes containing 0.5 ml of suspension. Unless otherwise stated, all samples were run at 37°C and preheated in a water bath before mixing. For the more rapidly transported species, it was found that preheating the sample was sufficient to permit temperature equilibration within l5s and measurement could be started within 30s. To prevent sedimentation in low haematocrit samples, air bubbles were passed through the sample tube between data accumulations. The air was administered through a fine plastic tube from a peristaltic pump and the bubble rate monitored by the temporary shift in the lock signal each time a bubble passed through the receiving coil. This was satisfactory for haematocrits below 70 %, but was not suitable or necessary for very viscous high-haematocrit samples. The compounds used to enhance the differential magnetic susceptibility included Cationised Ferritin (Miles Laboratories, Stoke Poges, Slough, U.K.) and AnalaR chlorides of Fe3+, Mn2+ and Dy3+ (BDH Chemicals, Poole, Dorset, U.K.). These metal ions were liganded to desferrioxamine (CIBA Laboratories, Horsham, West Sussex, U.K.), EDTA or 'diethylenetriamine penta-acetic acid' (NN-bis-'2[bis(carboxymethyl)amino]ethyl}glycine, DTPA) (BDH). D- and L-Alanine and L-lactate were obtained from BDH.
Physical Principles of the Methods Used Spili-echoes The n.m.r. spectra in this paper are obtained using spin-echo methods followed by Fourier transformation. The collection of spectra using a two-pulse spin-echo sequence has the advantage that resonances with relatively long values of T2 can be selected from a spectrum (Campbell et al., 1975; Brown et al., 1977). The simple two-pulse sequence 90'-T- 180'
K. M. BRINDLE AND OTHERS
produces an 'echo' at time 2T after the 900 pulse (Carr & Purcell, 1954), with amplitude given by: S(2r) = S(0)ep(~2r 2Dy2G2T3}F()
(1)
where y is the magnetogyric ratio, G is the magnetic field gradient across the sample and D is the diffusion coefficient of the observed molecule in the solution. The F(J) term leads to a modulation of the signal if there is homonuclear spin-spin coupling in the system. For a singlet F(J) = 1 at all times, but for a first-order doublet F(J) = cos(2nJT) (Freeman & Hill, 1975). This makes it a useful assignment aid. For example, the alanine methyl resonance is coupled to the a-CH with coupling constant J = 7.3 Hz, thus when 2r = 136ms the methyl resonance is inverted because F(J) = -1. The physical meaning of the term involving 93, G2 and D is that the echo is not properly refocused if a molecule diffuses to a region of different applied field during the time required to produce the echo. In a spectrometer with good homogeneity and with a homogeneous sample, this term is negligible for T values of up to at least lOOms. However, at long values of t, or when G is'large, this term makes the amplitude of the echo decay rapidly (Abragam, 1961) and with calibrated applied field gradients this is a good method for measuring diffusion coefficients (Stejskal & Tanner, 1965). Andrasko (1976) has measured Li+ transport into erythrocytes using a method that utilizes the fact that the diffusion of molecules inside a cell is more restricted than outside. Pulsed field gradients were applied to the sample in a way which allowed the fraction of Li+ ions inside the cell to be measured. The effects of the G2D term can be removed by applying many 180° pulses between the 90° pulse and data collection since r is then kept short as far as this term is concerned. A suitable pulse sequence is the Carr-Purcell-Meiboom-Gill sequence with r 1 ms (Meiboom & Gill, 1958; Freeman & Hill, 1975). With this sequence the decay rate is dominated by the T2 term. Fig. I demonstrates the behaviour of the echo amplitude for extracellular glycine as a function of time in three different situations. In 2H20 Krebs/ Ringer buffer the amplitude of the a-CH2 resonance decays at a rate that corresponds to a half-life of 1.5s with both the simple two-pulse sequence and the multiple-pulse sequence. In an 80% suspension of erythrocytes, however, the observed decay rate is much faster with the simple sequence (half-life 25 ms) than with the multiple-pulse sequence (half-life 280ms). In this experiment all the glycine is effectively outside the cell, because the transport rate is slow and the experiments were done relatively quickly after the addition of glycine. Glycine is also convenient because the a-CH2 resonance is a singlet and F(J) = 1 1979
39
MEMBRANE TRANSPORT STUDIED BY N.M.R. 17
=___
(a)
0
11
\-%~000 O..0--.O 0
7
(b)
2
(c)
0
0.2
0.1
0.3
Time interval (s) Fig. 1. Effect of diffusion on spin-echo anmplitude The height of the CH2 resonance of glycine observed in spin-echo experiments as a function of the time interval between the 900 pulse and the start of data acquisition. The experiment is arranged so that essentially all the glycine is outside the erythrocyte. Other conditions were: temperature, 293 K; haematocrit, 84%Y; [glycine]00,, 75 mM. o, Simple 900-r180°-r sequence on the suspension of cells ([DyDTPA]OUt 0.15mM); v, multiple-pulse sequence (Carr-Purcell-Meiboom-Gill) on the same sample; EO, a control experiment on a cell-free solution of glycine (12mM) and Dy-DTPA (0.15mM) using a 90 -r-180'-T sequence.
in eqn. (1). A dysprosium complex has also been added to the solutions (see below), but the ratio of glycine to this complex has been kept constant by making allowance for the volume taken up by the cells. These results suggest that, if the simple twopulse sequence is used, the term involving 3 dominates the decay rate for molecules in the extracellular space. This, as we will show, is due to field gradients in the extracellular space.
Origin offield gradients In a suspension of particles, magnetic-susceptibility differences between the particles and the suspending medium can give rise to large field gradients. In the case of spheres these gradients arise only outside the spheres (see Fig. 2). According to Glasel & Lee (1 974), who performed experiments with glass beads, these gradients can be represented by the approximate equation: C= K(Bo) * Ax rl I[(r' + ror + r2)r (2) where K is a constant, Bo is the applied magnetic field, Ax is the difference in magnetic susceptibility Vol. 180 -
(d)
Fig. 2. Magnetic field in particle suspensions Illustration of the effects of different magnetic susceptibility on the lines of magnetic flux in different geometries: (a) Xi,,
